Non-line-of-sight object location estimation from scattered light using plenoptic data

Abstract

We investigate the use of plenoptic data for locating non-line-of-sight (NLOS) objects from a scattered light signature. Using Fourier analysis, the resolution limits of the depth and transversal location estimates are derived from fundamental considerations on scattering physics and measurement noise. Based on the refocusing algorithm developed in the computer vision field, we derive an alternative formulation of the projection slice theorem in a form directly connecting the light field and a full spatial frequency spectrum including both depth and transversal dimensions. Using this alternative formulation, we propose an efficient spatial frequency filtering method for location estimation that is defined on a newly introduced mixed space frequency plane and achieves the theoretically limited depth resolution. A comparison with experimental results is reported.

Document Details

Document Type
Pub Defense Publication
Publication Date
Jan 22, 2021
Source ID
10.1364/josaa.394846

Entities

People

  • James Leger
  • Takahiro Sasaki

Organizations

  • Defense Advanced Research Projects Agency

Tags

Fields of Study

  • Physics

Readers

  • Image Processing and Computer Vision.
  • Sensor Fusion and Tracking Systems.
  • Statistical inference.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms
  • Space